Histogram-equalization-based adaptive image sensor for real-time vision

Image sensing and coding are essential to vision machines. This paper presents an adaptive image sensor by using in situ histogram equalization. With this histogram-equalization-based adaptation, this image sensor can adapt automatically to different lighting conditions without frame delay and provides a constant image signal excursion range. A CMOS prototype chip of 64/spl times/64-pixel is presented where each pixel contains ten transistors and occupies 50/spl times/50 /spl mu/m/sup 2/ in a classic 1-/spl mu/m double-metal single-poly process. The experimental results confirm an adaptation over illumination variation of 1:5000.